Bacharelado em Ciência da Computação (Sede)
URI permanente desta comunidadehttps://arandu.ufrpe.br/handle/123456789/6
Siglas das Coleções:
APP - Artigo Publicado em Periódico
TAE - Trabalho Apresentado em Evento
TCC - Trabalho de Conclusão de Curso
Navegar
2 resultados
Resultados da Pesquisa
Item The digital behavior of voters in interactions with the social media posts of candidates running for elections(2022-10-06) Silva Filho, Heriberto Alexandre da; Brito, Kellyton dos Santos; http://lattes.cnpq.br/8750956715158540; http://lattes.cnpq.br/6181814500468590The extensive use of digital tools and digital marketing strategies over the last few years has become increasingly more frequent and characteristic in political campaigns Within this scenario, this study aims to investigate the use of SM in contemporary political communication, seeking to understand the features that influence the engagement of voters in posts by politicians on their social media profiles. As a case study, we have focused on the Brazilian presidential election in 2018. The investigation was based on an analysis of politicians’ posts on Instagram, Twitter, and Facebook (N = 1319) in the last two weeks before the elections, which investigated features such as functional approach, the Aristotelian rhetoric adopted, and the type of content, among others, and established relationships between these features and user engagement. This study also proposes to investigate the feasibility of using machine learning models to predict the level of engagement of the candidate's posts. Finally, another objective of this paper is to find similarities or differences between the digital campaign strategies, and their impacts on the level of engagement, of the two candidates with the best electoral results. Our main results indicate that the platform with the highest level of engagement was Instagram, together with polarized discourses that presented speeches of attack and defense or emotionally charged topics tended to engage more. Regarding the predictions, the Gradient Boosting model proved to be efficient, R² =0.77, to make the predictions. Regarding the digital campaign strategies, although the two candidates are from opposite political sides, it was possible to find more similarities, such as: functional approach, content structure, and content type, and others...than differences. However the few differences found also represent a valuable result for the understanding of the political landscape, there were divergences for example in Aristotelian rhetoric, content type, and rhetorical device. All these results helped to understand how the electorate interacts with the candidates' speeches in a new era of digital campaigning.Item Métodos computacionais para a análise de dados de expressão gênica provenientes de uma análise de microarray utilizada para teste farmacológico(2023-04-28) Costa, Allan Mesquita da; Melo, Jeane Cecília Bezerra de; Costa, Luciana Amaral de Mascena; http://lattes.cnpq.br/2352032088330896; http://lattes.cnpq.br/8499459630583005; http://lattes.cnpq.br/2703136397519338The advent of the Human Genome Project (HGP), completed in October 2003, propelled the development of techniques for obtaining and analyzing biological data. The need to manage the vast amount of digital genome data was a decisive factor in the growth of a multidisciplinary area of knowledge, Computational Biology. In the two decades following the completion of the HGP, genomes of different organisms were obtained. Regarding mammals, projects such as the 1000 Genomes Project and the Cancer Genome Atlas (TCGA) illustrated the advancement of knowledge in the analysis of complex data. Among the newest techniques, we highlight Microarrays. They provide a significant amount of data in a single experiment, allowing the comparison of complete genomes. The analysis of Microarray data is relatively complex and requires protocols that make this analysis simpler, producing more understandable information. The present study involves the use of computational methods to analyze gene expression data obtained from a Microarray experiment used for pharmacological testing related to breast cancer. To process the raw data, obtained from a spreadsheet containing more than 3216 genes resulting from a Microarray analysis, a script was developed to facilitate the extraction of information from this data and subsequent selection of genes of interest. The program allowed the search for genes involved in the processes of cell death (apoptosis, necrosis, and autophagy), which are determining factors in the success analysis of the tested drug. To categorize the genes involved in the apoptotic, necrotic, and autophagic death cascade, heatmaps were constructed from fold-change values (difference in gene expression for values before and after treatment of cancerous cells with the mesoionic compound), using k-means clustering and hierarchical clustering techniques provided in the Heatmapper program. Results of the study include the development of a script in the R program that resulted in the separation of 20 genes involved in the apoptotic death cascade, six involved in the autophagic death, and seven involved in the necrotic death cascade, as well as the development of three heatmaps, contributing to the biological analysis of data, in addition to making Microarray data processing more accessible.